Patroller positioning method based on spliced images

Abstract

The invention discloses a patroller positioning method based on spliced images. Sequent images are shoot at certain angles in front of a preceding site and a current site by means of a binocular three-dimensional camera; the sequent images shoot by a left camera lens at the preceding site and the current site are spliced through a cylindrical projection method to obtain wide-angle spliced images in front of the left camera lens at the preceding site and the current site; the overlapped area of the spliced images of the left camera lens at the preceding site and the current site are calculated, Affine-SIFT feature extraction and matching are carried out on the overlapped area, and image matching feature points of the preceding site and the current site are looked for; the matching feature points are calculated reversely into an original image of the left camera lens, corresponding matching feature points are extracted in an image of a right camera lens, and faulted feature points are eliminated; the position and the posture of a patroller on the current site are solved in a light beam method. According to the technical scheme, high-precision posture determining and positioning of the patroller can be achieved.

Description

technical field [0001] The invention relates to the technical field of positioning a lunar surface patrol device, in particular to a method for positioning a patrol device based on spliced ​​images. Background technique [0002] The lunar surface patrol device is a mobile robot that performs close-range scientific exploration tasks in the complex and unstructured lunar surface environment. It can autonomously patrol the lunar surface or operate within hundreds of meters or even hundreds of kilometers of the lunar surface according to the command of the ground teleoperation center. , Realize the detection of the lunar surface environment through the scientific instruments carried by itself. In the process of autonomous cruising or remote control driving, the rover needs to know its own position and direction in the lunar surface coordinate system in order to unify its reconstructed lunar surface terrain under the same benchmark and guide subsequent path planning and travel. ...

AI Technical Summary

Problems solved by technology

However, in the lunar surface environment, the texture is relatively scarce and the lighting conditions are extremely complex, and the overlapping areas of the images taken by the rover at different positions are relatively limited and there are large scale differences

The lack of texture makes it difficult to extract feature points, and the complex lighting conditions and large-scale differences in images make the matching of image feature points prone to errors; affected by the above two factors, when performing feature point extraction and matching on limited overlapping areas of images, it may be When the number of feature points is too small, the error rate of feature point matching is high, or even the feature points cannot be extracted, it is impossible to obtain enough feature point information, and it is difficult to ensure the success of the patroller positioning

Images